Technology for inspection for detecting a defect of a printed matter using machine logic informed by machine learning. Some embodiments of the present invention may include one, or more, of the following features: (i) generates defect datasets; (ii) generates defect libraries; (iii) uses the generated defect libraries for deep learning training; and (iv) uses machine learning to detect defects using computer code (for example, a *.jpg format file) corresponding to an image of a piece of printed matter instead of using a visual image (that is, an image of the type that is created when a person takes a picture using a traditional film camera).
Legal claims defining the scope of protection. Each claim is shown in both the original legal language and a plain English translation.
2. The CIM of claim 1 wherein the defects library includes a plurality of historical data sets, with each historical data set including: (i) bitmap data corresponding to a bitmap of an original visual image used to make a printed image; (ii) bitmap data corresponding to a printed image of the original visual image; (iii) an indication of defect location(s) within the bitmap data corresponding to a printed image; and (iv) an indication of defect type defect location(s).
This invention relates to a computer-implemented method (CIM) for analyzing and managing defects in printed images. The system addresses the challenge of identifying and categorizing defects in printed materials by leveraging historical data to improve defect detection accuracy and efficiency. The method utilizes a defects library containing multiple historical data sets, each comprising four key components. First, it includes bitmap data representing the original visual image intended for printing. Second, it stores bitmap data of the actual printed image produced from the original. Third, it records defect locations within the printed image bitmap, pinpointing where imperfections occurred. Fourth, it categorizes these defects by type, providing detailed information on the nature of each defect. By comparing the original and printed bitmaps, the system identifies discrepancies and maps defect locations. The defect type classification further enhances analysis, enabling pattern recognition and root cause determination. This historical data allows for continuous improvement in defect detection algorithms, reducing errors in future print runs. The approach is particularly useful in industries requiring high-quality printed materials, such as manufacturing, packaging, and publishing, where defect detection is critical for maintaining product integrity and reducing waste.
3. The CIM of claim 1 wherein the defect types include at least one of the following: missing words, broken words, skip, scratch, blot, skew and/or extra ink.
This invention relates to a computer-implemented method for detecting and classifying defects in printed documents, particularly in printed text. The problem addressed is the need for automated systems to accurately identify and categorize various types of printing defects that can affect document quality, such as missing words, broken words, skips, scratches, blots, skew, and extra ink. These defects can arise from mechanical issues in printing equipment, ink problems, or misalignment, and their detection is crucial for quality control in industries like publishing, packaging, and document processing. The method involves analyzing a printed document to detect and classify defects based on predefined defect types. The system scans the document and identifies anomalies in the printed text or images, comparing them against expected patterns or templates. Defects are classified into specific categories, such as missing words (where text is entirely absent), broken words (where parts of words are missing), skips (where printing is interrupted), scratches (surface marks), blots (ink smears), skew (misalignment of text or images), and extra ink (unintended ink deposits). The classification helps in diagnosing the root cause of the defect, whether it is due to hardware malfunctions, ink issues, or other factors. This automated approach improves efficiency in quality control processes by reducing the need for manual inspection and providing detailed defect reports for corrective actions.
4. The CIM of claim 1 wherein the defect types include all of the following: missing words, broken words, skip, scratch, blot, skew and/or extra ink.
This invention relates to a computer-implemented method for detecting and classifying defects in printed documents. The method addresses the challenge of identifying various types of printing errors that can occur during document production, such as missing words, broken words, skips, scratches, blots, skew, and extra ink. These defects can degrade print quality, affect readability, and lead to production waste if undetected. The method involves analyzing a printed document to identify and categorize defects based on predefined defect types. The system captures an image of the printed document and processes it to detect anomalies in text and visual elements. For missing words, the system checks for gaps where text should appear. Broken words are identified by analyzing partial or fragmented characters. Skips are detected by comparing expected text positions with actual printed content. Scratches and blots are identified through image processing techniques that recognize irregular marks. Skew is detected by analyzing the alignment of text lines relative to a reference. Extra ink is identified by detecting unintended marks or smudges outside the expected text or image areas. The method may also include generating a report summarizing the detected defects, their locations, and severity levels. This allows for quality control and corrective actions in printing processes. The system can be integrated into automated printing workflows to ensure consistent output quality. The invention improves efficiency in print production by reducing manual inspection efforts and minimizing waste from defective prints.
5. The CIM of claim 1 wherein the ML algorithm is a collection of machine logic based rules that evolve over time as new training data is used to refine and improve effectiveness and reliability of the ML algorithm.
This invention relates to a machine learning (ML) algorithm used within a cybersecurity information management (CIM) system. The problem addressed is the static nature of traditional rule-based security systems, which struggle to adapt to evolving threats. The solution involves an ML algorithm that dynamically evolves by refining its logic-based rules over time as new training data is incorporated. This adaptive approach improves the algorithm's effectiveness and reliability in detecting and responding to cybersecurity threats. The system continuously learns from new data, allowing it to enhance its decision-making processes without manual intervention. The evolving rules ensure that the algorithm remains effective against emerging threats, reducing false positives and improving threat detection accuracy. This adaptive learning mechanism distinguishes the invention from static rule-based systems, providing a more resilient and scalable security solution. The invention is particularly useful in environments where threat landscapes change rapidly, such as enterprise networks or cloud-based systems. By leveraging continuous learning, the ML algorithm maintains high performance without requiring frequent manual updates.
8. The CPP of claim 7 wherein the defects library includes a plurality of historical data sets, with each historical data set including: (i) bitmap data corresponding to a bitmap of an original visual image used to make a printed image; (ii) bitmap data corresponding to a printed image of the original visual image; (iii) an indication of defect location(s) within the bitmap data corresponding to a printed image; and (iv) an indication of defect type defect location(s).
This invention relates to defect detection and analysis in printed images, addressing the challenge of identifying and categorizing defects in printed outputs. The system uses a defects library containing historical data sets to improve defect detection accuracy. Each data set includes bitmap data of an original visual image, bitmap data of the printed image, defect locations within the printed image, and defect types associated with those locations. The library enables comparison between original and printed images to identify discrepancies, such as misprints, color variations, or alignment errors. By analyzing historical data, the system can learn patterns and improve future defect detection. The defects library supports automated quality control in printing processes, reducing manual inspection and enhancing consistency. The invention is particularly useful in high-volume printing environments where maintaining print quality is critical. The system leverages bitmap data to precisely locate and classify defects, ensuring accurate and efficient defect analysis.
9. The CPP of claim 7 wherein the defect types include at least one of the following: missing words, broken words, skip, scratch, blot, skew and/or extra ink.
This invention relates to a content processing platform (CPP) designed to detect and correct defects in printed or scanned documents. The system addresses the challenge of identifying and rectifying various types of defects that degrade document quality, such as missing words, broken words, skips, scratches, blots, skew, and extra ink. These defects can arise from printing errors, scanning issues, or physical damage to documents, leading to inaccuracies in text recognition and data extraction. The CPP employs advanced image processing and pattern recognition techniques to analyze documents and classify defects based on their type. For missing words, the system identifies gaps in text where content is entirely absent. Broken words are detected when parts of a word are illegible or separated. Skips occur when lines or sections of text are omitted, while scratches and blots are physical marks that obscure text. Skew refers to misalignment in document orientation, and extra ink appears as unintended marks or smudges. Once defects are identified, the CPP applies corrective measures such as interpolation, pattern matching, or contextual analysis to restore the original content. For example, missing words may be reconstructed using surrounding text context, while skew can be corrected through alignment algorithms. The system ensures that the final output is accurate and usable for further processing, such as optical character recognition (OCR) or data extraction. This technology is particularly useful in industries like document archiving, legal, and financial services where document integrity is critical.
10. The CPP of claim 7 wherein the defect types include all of the following: missing words, broken words, skip, scratch, blot, skew and/or extra ink.
The invention relates to a system for detecting and classifying print defects in printed documents. The system addresses the challenge of identifying various types of defects that can occur during the printing process, which may affect the readability and quality of printed materials. These defects can include missing words, broken words, skips, scratches, blots, skew, and extra ink, among others. The system is designed to analyze printed documents and automatically detect these defect types to ensure high-quality output. The system employs image processing techniques to scan and evaluate printed documents for defects. It uses pattern recognition and machine learning algorithms to identify and classify different types of defects based on their visual characteristics. For example, missing words are detected when expected text is absent, while broken words are identified when parts of a word are incomplete. Skips occur when sections of text are omitted, and scratches are thin, linear marks that disrupt the printed surface. Blots are irregular ink marks, skew refers to misalignment in the printed text, and extra ink appears as unintended ink deposits. By categorizing these defects, the system helps in diagnosing printing equipment issues, improving print quality control, and reducing waste. The ability to detect multiple defect types ensures comprehensive quality assurance, making the system suitable for high-volume printing environments such as publishing, packaging, and document production. The system may be integrated into existing printing workflows to provide real-time defect detection and reporting.
11. The CPP of claim 7 wherein the ML algorithm is a collection of machine logic based rules that evolve over time as new training data is used to refine and improve effectiveness and reliability of the ML algorithm.
This invention relates to a machine learning (ML) algorithm used in a cyber-physical process (CPP) system, addressing the challenge of maintaining and improving the effectiveness and reliability of ML algorithms over time. The system employs a collection of machine logic-based rules that dynamically evolve as new training data is incorporated. These rules are refined iteratively to enhance the algorithm's performance, ensuring it adapts to changing conditions and improves accuracy. The ML algorithm operates within a CPP system, which integrates computational and physical processes to monitor, control, or optimize operations in real-world environments. The evolving nature of the rules allows the algorithm to learn from new data, reducing errors and increasing reliability without requiring manual updates. This approach is particularly useful in applications where environmental factors or operational conditions change frequently, such as industrial automation, smart infrastructure, or autonomous systems. The system ensures continuous improvement by leveraging feedback loops and adaptive learning mechanisms, making it more robust and efficient over time.
14. The CS of claim 13 wherein the defects library includes a plurality of historical data sets, with each historical data set including: (i) bitmap data corresponding to a bitmap of an original visual image used to make a printed image; (ii) bitmap data corresponding to a printed image of the original visual image; (iii) an indication of defect location(s) within the bitmap data corresponding to a printed image; and (iv) an indication of defect type defect location(s).
This invention relates to defect detection and analysis in printed images. The technology addresses the challenge of identifying and classifying defects in printed materials by leveraging historical data to improve accuracy and efficiency in defect detection systems. The system includes a defects library containing multiple historical data sets, each of which stores detailed information about defects found in printed images. Each data set includes bitmap data representing the original visual image used to create the printed image, as well as bitmap data of the printed image itself. Additionally, the data set records the locations of defects within the printed image bitmap and specifies the type of each defect. This structured approach allows for comprehensive defect tracking, enabling the system to compare new printed images against historical defect patterns to identify and classify defects more effectively. By analyzing past defect occurrences, the system can improve defect detection algorithms, reduce false positives, and enhance overall print quality control. The historical data sets serve as a reference for training machine learning models or refining rule-based detection methods, ensuring that the system adapts to recurring defect types and locations. This method supports automated quality assurance in printing processes, reducing manual inspection efforts and improving production efficiency.
15. The CS of claim 13 wherein the defect types include at least one of the following: missing words, broken words, skip, scratch, blot, skew and/or extra ink.
The invention relates to a system for detecting and classifying defects in printed documents, particularly focusing on identifying specific types of printing errors. The system analyzes printed documents to detect various defect types, including missing words, broken words, skips, scratches, blots, skew, and extra ink. These defects can occur due to mechanical issues in printing equipment, ink problems, or misalignment during the printing process. The system processes the printed document to identify these defects, ensuring that the final output meets quality standards. By categorizing the defects, the system helps in diagnosing the root cause of printing errors, allowing for targeted maintenance and adjustments to the printing machinery. This improves overall print quality and reduces waste by identifying and correcting defects before the documents are distributed. The system may use image processing techniques, such as optical character recognition (OCR) and pattern recognition, to analyze the printed output and compare it against a reference or expected output. The classification of defects enables automated quality control, reducing the need for manual inspection and increasing efficiency in large-scale printing operations.
16. The CS of claim 13 wherein the defect types include all of the following: missing words, broken words, skip, scratch, blot, skew and/or extra ink.
This invention relates to a system for detecting and classifying defects in printed documents, addressing the challenge of identifying various types of printing errors that can affect document quality and readability. The system analyzes printed documents to detect defects such as missing words, broken words, skips, scratches, blots, skew, and extra ink. These defects are categorized to enable targeted quality control and correction processes. The system may use image processing techniques, such as optical character recognition (OCR) and pattern recognition, to identify and classify the defects based on their visual characteristics. By detecting these specific defect types, the system helps improve printing accuracy and reduce waste in document production. The invention may be integrated into printing workflows to automatically flag defective prints for review or reprocessing, ensuring higher-quality output. The system's ability to recognize multiple defect types allows for comprehensive defect detection, enhancing overall printing efficiency and reliability.
17. The CS of claim 13 wherein the ML algorithm is a collection of machine logic based rules that evolve over time as new training data is used to refine and improve effectiveness and reliability of the ML algorithm.
This invention relates to a machine learning (ML) system designed to improve its effectiveness and reliability over time. The system uses a collection of machine logic-based rules that are dynamically refined as new training data becomes available. The ML algorithm continuously evolves by incorporating updated data, allowing it to adapt to changing conditions and improve its performance. This adaptive approach ensures that the system remains accurate and reliable as it processes new information. The invention addresses the challenge of maintaining ML model accuracy in dynamic environments where data patterns may shift over time. By continuously refining its rules, the system avoids degradation in performance and ensures consistent results. The adaptive ML algorithm is particularly useful in applications where real-time decision-making is critical, such as predictive analytics, fraud detection, or autonomous systems. The system's ability to learn from new data without manual intervention reduces the need for frequent model retraining, making it more efficient and scalable. The invention provides a self-improving ML framework that enhances reliability and effectiveness in real-world applications.
Cooperative Patent Classification codes for this invention. Click any code to explore related patents in that topic.
June 30, 2021
April 23, 2024
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.